Recent research has indicated high rates of substance abuse among adolescents with emotional and behavioral disorders. Moreover, adolescents in clinical and correctional settings found to have comorbid disorders involving substance abuse experience higher morbidity and mortality rates when compared to adolescents having one or no condition. The present study examines the ability of the Problem-Oriented Screening Instrument for Teenagers (POSIT) to identify DSM-III-R-defined psychoactive substance use disorders among 342 adolescents aged 12-19 years. Participants were sampled from school, clinical, and correctional settings. Optimal-scale cut scores for drug abuse diagnosis classification were derived by a minimum loss function method that minimized false classifications. When using the optimal cut score of two for die total sample, the standard POSIT substance use/abuse scale obtained a drug abuse diagnosis classification accuracy of 84% with sensitivity and specificity ratios of 95% and 79%, respectively. The internal validity of the standard 17-item substance use/abuse scale was subsequently examined by principle component analysis, item analysis, and coefficient alpha. The internal validity analyses were conducted to determine if a shortened scale could be developed and yet retain acceptable classification accuracy. When using the optimal cut score of two for the total sample, the revised 11-item scale obtained a drug abuse diagnosis classification accuracy of 85% with sensitivity and specificity ratios of 91% and 82%, respectively. The results suggest with the POSIT can serve as a useful first-gate instrument to identify adolescents in need of further drug abuse assessment.
|Original language||English (US)|
|Number of pages||20|
|Journal||American Journal of Drug and Alcohol Abuse|
|State||Published - 1997|
Bibliographical noteFunding Information:
This research was supported by a Scientist Development Award from the National Institute on Drug Abuse (K21 DA00254-01) to William W. Latimer and by NIDA grant DA05104. The authors wish to thank Elizabeth Rahdert, Ph.D., for guidance that helped to make the data analyses clinically useful and for manuscript review, and Nadav Cassuto for statistical consultation.
- Drug abuse
- Screening test